INDEX

Explanations

phrases that convey distancing or separation from others or concepts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

swick

-0.68

chance

-0.65

rano

-0.64

iop

-0.63

orable

-0.63

nosis

-0.63

opus

-0.60

ores

-0.59

frey

-0.59

amac

-0.58

POSITIVE LOGITS

 oneself

0.95

 themselves

0.83

 herself

0.82

 myself

0.82

 himself

0.80

 ourselves

0.80

iates

0.78

 yourselves

0.75

iveness

0.73

 yourself

0.72

Activations Density 0.012%

phrases that convey distancing or separation from others or concepts

No Comments

No Known Activations